[
https://issues.apache.org/jira/browse/HUDI-8812?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17935220#comment-17935220
]
Sagar Sumit commented on HUDI-8812:
-----------------------------------
Currently, columns being indexed in colstats (and by extension partition_stats
too) are being tracked in index definition. So, when user adds a column stats
are being colelcted automatically, so adding sql functionality is not a
blocker. Downgrading from Blocker to Major. Still, would be nice-to-have a sql
command to generate stats.
> Support Partition stats, column stats, bloom filters creation using SQL
> -----------------------------------------------------------------------
>
> Key: HUDI-8812
> URL: https://issues.apache.org/jira/browse/HUDI-8812
> Project: Apache Hudi
> Issue Type: Sub-task
> Components: index
> Reporter: Lokesh Jain
> Assignee: Sagar Sumit
> Priority: Blocker
> Fix For: 1.0.2
>
> Original Estimate: 12h
> Remaining Estimate: 12h
>
> Users cannot create partition_stats, column_statas and bloom_filters indexes
> through SQL. They can only be created through datasource. There are
> workarounds e.g. one can create expression index using column_stats with
> identity function of a column. But, we should provide more standard sql for
> building these indexes.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
